Structure and Mechanism of Pseudomonas aeruginosa PA0254/HudA, a prFMN-Dependent Pyrrole-2-carboxylic Acid Decarboxylase Linked to Virulence

Article abstract of DOI:10.1021/acscatal.0c05042

The UbiD family of reversible (de)carboxylases depends on the recently discovered prenylated-FMN (prFMN) cofactor for activity. The model enzyme ferulic acid decarboxylase (Fdc1) decarboxylates unsaturated aliphatic acids via a reversible 1,3-cycloaddition process. Protein engineering has extended the Fdc1 substrate range to include (hetero)aromatic acids, although catalytic rates remain poor. This raises the question how efficient decarboxylation of (hetero)aromatic acids is achieved by other UbiD family members. Here, we show that the Pseudomonas aeruginosa virulence attenuation factor PA0254/HudA is a pyrrole-2-carboxylic acid decarboxylase. The crystal structure of the enzyme in the presence of the reversible inhibitor imidazole reveals a covalent prFMN-imidazole adduct is formed. Substrate screening reveals HudA and selected active site variants can accept a modest range of heteroaromatic compounds, including thiophene-2-carboxylic acid. Together with computational studies, our data suggests prFMN covalent catalysis occurs via electrophilic aromatic substitution and links HudA activity with the inhibitory effects of pyrrole-2-carboxylic acid on P. aeruginosa quorum sensing.

Full text of DOI:10.1021/acscatal.0c05042

pubs.acs.org/acscatalysis
Research Article
Structure and Mechanism of Pseudomonas aeruginosa PA0254/
HudA, a prFMN-Dependent Pyrrole-2-carboxylic Acid Decarboxylase
Linked to Virulence
Karl A. P. Payne, Stephen A. Marshall, Karl Fisher, Stephen E. J. Rigby, Matthew J. Cliff, Reynard Spiess,
Diego M. Cannas, Igor Larrosa, Sam Hay, and David Leys*
Cite This: ACS Catal. 2021, 11, 2865−2878
Read Online
ACCESS
Metrics & More
Article Recommendations
*
sı Supporting Information
ABSTRACT: The UbiD family of reversible (de)carboxylases
depends on the recently discovered prenylated-FMN (prFMN)
cofactor for activity. The model enzyme ferulic acid decarboxylase
(
1
Fdc1) decarboxylates unsaturated aliphatic acids via a reversible
,3-cycloaddition process. Protein engineering has extended the
Fdc1 substrate range to include (hetero)aromatic acids, although
catalytic rates remain poor. This raises the question how efficient
decarboxylation of (hetero)aromatic acids is achieved by other
UbiD family members. Here, we show that the Pseudomonas
aeruginosa virulence attenuation factor PA0254/HudA is a pyrrole-
2
-carboxylic acid decarboxylase. The crystal structure of the
enzyme in the presence of the reversible inhibitor imidazole reveals
a covalent prFMN−imidazole adduct is formed. Substrate
screening reveals HudA and selected active site variants can accept a modest range of heteroaromatic compounds, including
thiophene-2-carboxylic acid. Together with computational studies, our data suggests prFMN covalent catalysis occurs via
electrophilic aromatic substitution and links HudA activity with the inhibitory effects of pyrrole-2-carboxylic acid on P. aeruginosa
quorum sensing.
KEYWORDS: decarboxylase, enzyme mechanism, flavin chemistry, prFMN, Pseudomonas aeruginosa, quorum sensing,
pyrrole-2-carboxylic acid
1
1
INTRODUCTION
catalytic cycle. However, the extent to which 1,3-dipolar
cycloaddition occurs for UbiD enzymes acting on (hetero)-
aromatic substrates has been questioned, as the necessary
dearomatization of the substrate presents a substantial barrier
to cycloadduct formation.
■
The UbiD family of enzymes catalyzes the reversible
nonoxidative decarboxylation of a wide range of unsaturated
aliphatic and aromatic compounds, the latter including
1
2
3
phenolic compounds, heteroaromatics, phthalates, poly-
4
5
Unfortunately, recent structural and biochemical character-
ization of three UbiD decarboxylases acting on aromatic
substrates has not yielded detailed mechanistic insights to
answer this question. In the case of the canonical Escherichia
coli UbiD, the purified enzyme failed to mature the active form
of the cofactor, instead accumulating an inactive prFMN
species. Active enzymes could be obtained for HmfF
catalyzing the decarboxylation of 2,5-furandicarboxylic
cyclics, as well as benzene itself (recently reviewed in ref 6).
Recent insights into the UbiD mode of action came from
7
studies on the fungal enzyme ferulic acid decarboxylase Fdc1.
These revealed that UbiD enzymes require a modified flavin
cofactor, prenylated-FMN (prFMN), for activity (Figure 1).
The genetically associated UbiX acts as the flavin prenyl-
transferase, attaching a prenyl moiety to the N5 and C6
positions of reduced FMN, thereby extending the cofactor with
8
radical
1
2
(
1
3
9
,10
acid ) and AroY (decarboxylating protocatechuate to
a fourth nonaromatic ring.
reduced prFMNH produced by UbiX is proposed to undergo
Following UbiD binding, the
2
iminium
Received: November 18, 2020
Revised: February 4, 2021
Published: February 17, 2021
oxidative maturation to yield the active prFMN
The azomethine ylide character of the prFMN
reversible 1,3-dipolar cycloaddition underpinning the (de)-
carboxylase mechanism of Fdc1 (Figure 1). Recent structural
species.
supports a
iminium
8
insights into a range of covalently bound substrate/cofactor
adducts confirmed that cycloadducts are formed during the
©
2021 The Authors. Published by
American Chemical Society
https://dx.doi.org/10.1021/acscatal.0c05042
2
865
ACS Catal. 2021, 11, 2865−2878

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
ture-based semirational engineering supported a modest
improvement in yields with the latter compound. In
combination with the DFT computational studies, our data
suggests that covalent catalysis occurs via electrophilic
aromatic substitution.
EXPERIMENTAL PROCEDURES
Cloning of PA0254 Pyrrole-2-carboxylate Decarbox-
ylase. The gene encoding PA0254 (NC_002516) was codon
■
Table 1. Crystallographic Data Collection and Refinement
Parameters
PAO254/HudA
imidazole complex
PAO254/HudA N318A
FMN complex
PDB
7ABN
7ABO
resolution range
50−1.65 (1.68−1.65)
47.0−1.95 (1.98−1.95)
(
Å)
Figure 1. Covalent catalysis by prFMN in UbiD enzymes. prFMN
cofactor is formed through FMN prenylation followed by oxidative
space group
unit cell
P12 1
P12 1
1
1
107.81, 55.48, 199.53, 90, 107.88, 55.57, 198.99, 90,
maturation (prenylation shown in red). Unsaturated carboxylic acid is
proposed to form a covalent adduct with the prFMNiminium cofactor
99.93, 90
27 9939 (18 328)
3.1 (2.9)
96.4 (86.0)
8.6 (1.5)
0.093 (0.812)
1.0 (0.7)
0.201 (0.298)
0.235 (0.341)
0.022
100.0, 90
16 8873 (8399)
3.2 (3.0)
99.2 (99.1)
11.7 (1.4)
0.081 (1.025)
1.0 (0.5)
0.214 (0.327)
0.246 (0.354)
0.004
unique reflections
multiplicity
completeness (%)
mean I/sigma(I)
R meas
CC1/2
R work
R free
RMS bonds (Å)
RMS angles (deg)
Ramachandran
favored (%)
Ramachandran
allowed (%)
Ramachandran
outliers (%)
leading to decarboxylation. In the case of acrylic acid substrates,
substantial evidence supports a 1,3-dipolar cycloaddition process
leading to intermediate 1 (Int 1), followed by decarboxylation to form
Int 2 + CO . Exchange of CO with a conserved acidic residue leads
2
2
to protonation to form Int 3, which is proposed to form product
through cycloelimination. The exact nature of the various
intermediate species remains unclear in the case of (hetero)aromatic
substrates.
1
4
2.12
96.3
0.68
95.9
catechol ). In each case, crystals of the holo-enzyme did not
yield any substrate complexes despite several attempts. The
discrepancy between the closed, solvent-occluded conforma-
tion of the Fdc1 active site (which readily binds substrates or
inhibitors) and the more open, solvent-accessible conforma-
tions (which hitherto have not yielded any ligand-bound
structures) observed for the UbiD, HmfF, and AroY structures
can be explained by a putative hinge motion of the prFMN-
3.4
3.7
0.3
0.4
average B factor
19.0
38.0
2
(
Å )
1
4−16
binding domain.
Such a conformational change would
link the postulated open and closed conformations, but as yet
no UbiD enzyme has been demonstrated to exhibit both
conformations. To further complicate matters, the dimeric
Fdc1 belongs to a distinct branch of the UbiD family tree as
compared to the hexameric UbiD/HmfF and AroY enzymes,
reflected in the distinct substrate specificities of these enzymes.
However, recent protein engineering studies on Fdc1 have
been able to extend the substrate scope to include (hetero)-
optimized to remove codons that were rare in E. coli and
synthesized (Genscript). The gene was amplified using
Phusion polymerase (NEB) using the primers
P0254pNIC28F (TACTTCCAATCCATGAACCGCTCGG-
CACTG) and P0254pNIC28R (GTGCGGCCGCAAGCTT-
CAGGCATCACCAAAGCC), and the PCR product was
cloned into the NcoI and HindIII sites of pNic28-Bsa4
expression plasmid (Addgene) using Infusion HD (Clontech)
15
16
aromatic compounds using a single active site substitution.
to reproduce the construct reported by Jacewicz et al. Once
This suggests that quaternary structure or position within the
UbiD-family tree does not have a fundamental link to substrate
specificity.
the sequence of the desired insert was confirmed, the
corresponding purified plasmid was transformed into E. coli
BL21(DE3) with ubiXpET21b to provide sufficient levels of
prFMN in vivo. BL21(DE3) cells were also transformed with
PA0254pNic28 without ubiXpET21b.
Mutagenesis. Mutagenesis primers were designed using
the QuikChange Primer Design Program (http://www.
genomics.agilent.com/primerDesignProgram.jsp). PCR was
performed using Phusion polymerase (NEB). Template was
removed by DpnI (NEB) digest, and the PCR product was
transformed into E. coli NEB5α. Once the presence of the
desired mutation was confirmed by DNA sequencing, the
plasmid was cotransformed with ubiXpET21b into E. coli
BL21(DE3).
8
We sought to provide further detailed insights into the UbiD
reaction with heteroaromatic substrates, focusing on pyrrole-2-
carboxylate (P2C) decarboxylases. We here report that the
UbiD-like Pseudomonas aeruginosa virulence attenuation factor
PA0254/HudA is a close homologue of Bacillus megaterium
PYR2910 pyrrole-2-carboxylate (P2C) decarboxylase and show
that PA0254 is capable of prFMN-dependent decarboxylation
of P2C as well as carboxylation of pyrrole. Structure
determination of the PA0254 holo-enzyme reveals a closed
conformation with a buffer-derived imidazole moiety cova-
lently bound to the prFMN. Substrate screening reveals that a
modest range of heteroaromatic substrates is accepted,
including weak activity with thiophene-2-carboxylate. Struc-
Protein Expression. The enzymes were expressed in
BL21(DE3) grown at 37 °C/180 rpm in TB broth
2
866
https://dx.doi.org/10.1021/acscatal.0c05042
ACS Catal. 2021, 11, 2865−2878

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
supplemented with 50 μg/mL kanamycin and 50 μg/mL
ampicillin. At mid log phase cells were induced with 0.25 mM
and 20 μL was added to 980 μL of 50% v/v H O/acetonitrile.
2
Sample analysis was performed by HPLC as described above.
UbiX
IPTG and supplemented with 1 mM MnCl , grown overnight
PA0254
Activity Dependency on Metal Ions. To
2
UbiX
at 17 °C/180 rpm, and then harvested by centrifugation (4 °C,
test for the metal ion requirements of PA0254 , protein
purified in the presence of NaCl or KCl was reconstituted with
prFMN produced in vitro. Two UbiX reactions were
performed in parallel, with 200 mM NaCl or KCl, 50 mM
Tris, pH 7.5, as the diluent; 0.1 mM FMN, 0.5 mM DMAP, 0.5
mM NADH, 50 μM UbiX, and 2 μM Fre (E. coli NAD(P)H-
flavin reductase) were mixed with each reaction inside a Belle
Technologies Anaerobic chamber. The reactions were
7
000g for 10 min).
Purification of His-Tagged Proteins. Cell pellets were
resuspended in buffer A (200 mM KCl, 1 mM MnCl , 50 mM
2
Tris pH 7.5) supplemented with DNase, RNase, lysozyme
(
(
Sigma), and Complete EDTA-free protease inhibitor cocktail
Roche). Cells were lysed using a French press at 20 000 psi,
and the lysate was clarified by centrifugation at 125 000g for 90
min. The supernatant was applied to a Ni−NTA agarose
column (Qiagen). Initially, imidazole was used to elute the
protein with the column being washed with 3 column volumes
of buffer A supplemented with 10 mM imidazole and protein
eluted in 1 mL fractions with buffer A supplemented with 250
mM imidazole. Once imidazole was observed to be binding to
the cofactor, subsequent purifications utilized histidine to elute
the protein, with the column being washed with 3 column
volumes of buffer A supplemented with 10 mM histidine and
protein eluted in 1 mL fractions with buffer A supplemented
with 100 mM histidine. Samples were subjected to SDS-PAGE
analysis, and fractions found to contain the purified protein
were pooled. Imidazole/histidine was removed using a 10-DG
desalting column (Bio-Rad) equilibrated with buffer A. Protein
was aliquoted and flash frozen until required. Cells grown for
the production of apo PA0254 (i.e., without ubiXpET21b)
were resuspended in 200 mM NaCl, 50 mM Tris-HCl, pH 7.5
or 200 mM KCl, 50 mM Tris-HCl, pH 7.5. Other aspects of
the purification remained constant with holo preparations using
imidazole as the eluent.
incubated for 3 h prior to separation of prFMNH from the
proteins using a 10 kDa MWCO microcentrifugal concentrator
2
(
Sartorius). To a final concentration of 30 μM PA0254 in
either NaCl or KCl buffer A, filtrate containing prFMN was
added to a final concentration of 10 μM in the respective salt,
with MnCl or MgCl added to a final concentration of 100
2
2
μM. Controls with no MnCl /MgCl were performed as was a
2
2
control with no additional MnCl /MgCl , and prFMN. P2C
2
2
was dissolved in 50 mM NaCl, 50 mM NaPi, pH 6 or 50 mM
KCl, 50 mM KPi, pH 6. PA0254 in the respective salt was
added to a final concentration of 0.75 μM (final added
[prFMN] = 0.25 μM). HPLC assays were performed as
described above with the exception that reactions were allowed
to proceed for 1 h at 25 °C prior to quenching with acetonitrile
+
0.1% TFA in a 1:1 volume ratio with the reaction, followed
by centrifugation and a 1 in 5 dilution with 50% (v/v) H O/
2
acetonitrile + 0.1% TFA. Detection and depletion of P2C was
performed at 270 nm.
1H NMR Monitored Enzyme-Catalyzed Deuterium
Exchange. Ten millimolar substrate, 100 mM NaPi, pH 5.6,
in D O was incubated overnight at 30 °C with and without 5
2
UV−vis Spectroscopy/Protein Quantification and
Decarboxylation Assays. UV−vis absorbance spectra were
recorded with a Cary 50 Bio UV−Vis spectrophotometer
UbiX
μM PA0254 . Data were collected on a Bruker 500 MHz
NMR spectrometer and QCI-F cryoprobe at 298 K with a 4
min accumulation time.
EPR Spectroscopy. Continuous-wave X-band (∼9.4 GHz)
EPR spectra were recorded with a Bruker E500/580 EPR
spectrometer with a Bruker “Super High Q” cavity (ER
(
Varian). The protein concentration was estimated from the
A280 absorption peak with extinction coefficients calculated
from the primary amino acid sequence using the ProtParam
program on the ExPASy proteomics server. PA0254 concen-
4
122SHQE) coupled to an Oxford Instruments ESR900
tration was estimated using ε280 = 78 380 M− cm . Initial
rates of pyrrole-2-carboxylate (P2C) decarboxylation were
determined by monitoring P2C concentration by UV−vis
1
−1
helium flow cryostat for temperature control. Spectra were
collected at 20 K using 10 μW microwave power, 100 kHz field
modulation frequency, and 1 G modulation amplitude.
spectroscopy at 255 nm using an extinction coefficient ε
=
UbiX
UbiX
2
55
Crystallization of PA0254 . Purified PA0254
in
−
1
−1
1
8 000 M cm . Assays were performed with various
200 mM NaCl, 50 mM Tris, pH 7.5, was concentrated in a
concentrations of substrate dissolved in 350 μL of 50 mM
Vivaspin 30 kDa MWCO spin concentrator to a final
concentration of ∼10−20 mg/mL. Initial screening by sitting
drop was performed; mixing 0.3 μL of protein with 0.3 μL of
mother liquor led to crystals in a variety of conditions when
incubated at 21 °C. The best-performing crystals originated
from well D3 of the Morpheus commercial screen (Molecular
Dimensions) consisting of 0.12 M alcohols, 0.1 M imidazole/
KCl, 50 mM NaPi, pH 6 using a 1 mm path length cuvette.
UbiX
PA0254
Decarboxylation Reactions Assayed by
HPLC. Typical assays containing 10 mM pyrrole-2-carboxylate,
0 mM KCl, and 50 mM NaPi, pH 6, were incubated with and
5
without enzyme at 30 °C overnight. The sample was
centrifuged at 16 100g to remove precipitate, and 50 μL was
added to 450 μL of 50% v/v H O/acetonitrile. Sample analysis
2
MES, pH 6.5, 20% v/v glycerol, and 10% w/v PEG 4000.
UbiX
was performed using an Agilent 1260 Infinity Series HPLC
equipped with a UV detector. The stationary phase was a
Kinetex 5 μm C18 100A column, 250 × 4.6 mm. The mobile
phase was acetonitrile/water (50/50) with 0.1% TFA at a flow
rate of 1 mL/min, and unless otherwise stated, detection was
Crystals of PA0254
N318H mutant were attained as above
but were grown in condition H7 of the Morpheus screen
(Molecular Dimensions) consisting of 0.1 M amino acids, 0.1
M NaHEPES/MOPS buffer, pH 7.5, 20% v/v glycerol, and
10% w/v PEG 4000.
performed at a wavelength of 210 nm.
Diffraction Data Collection and Structure Elucida-
tion. Crystals were flash cooled in liquid nitrogen. Data were
collected at Diamond beamlines and subsequently handled
UbiX
PA0254
Carboxylation Reactions Assayed by
HPLC. Typical assays containing 50 mM pyrrole, 100 mM
17
KPi, pH6, and 1 M KHCO (final pH 7.5) were incubated
using the CCP4 suite. All data were reduced and scaled using
3
18
with and without PA0254 enzyme at 30 °C overnight. The
XDS. Interpretable maps were obtained following molecular
sample was centrifuged at 16 100g to remove the precipitate,
replacement with the available apo-PA0254 crystal structure
2
867
https://dx.doi.org/10.1021/acscatal.0c05042
ACS Catal. 2021, 11, 2865−2878

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
(
PDB code 4IP2). The initial model was iteratively rebuilt and
17
refined using Coot and REFMAC5. The final model was
refined using data extending to 1.65 Å and contains 4
monomers in the asymmetric unit. For final data collection and
refinement statistics, see Table 1.
DFT Calculations. DFT calculations of the prFMN−
substrate adducts were performed using Gaussian 09, revision
D.01 at the B3LYP/6-311++G(d,p) level of theory with the
D3 version of Grimme’s dispersion with Becke−Johnson
1
9
damping and a generic polarizable continuum model of
water. Additional details of the models are provided in the
Supporting Information.
RESULTS
■
Figure 3. UV−vis spectrum of PA0254. PA0254 expressed with
Pyrrole-2-carboxylate Decarboxylase Belongs to the
HudA Clade. The amino acid sequence of the B. megaterium
PYR2910 pyrrole-2-carboxylate decarboxylase was kindly
provided by Professor Yoshida of Gifu University. Phylogenetic
analysis of the PYR2910 sequence indicates that it clusters with
(blue) and without (yellow) UbiX and after treatment with
NaBH CN under aerobic conditions (purple). Spectra were
3
normalized with respect to the A280 peak. (Inset) Close up of the
cofactor-related spectral features in the 300−800 nm range.
Figure 2. Phylogenetic analysis of known UbiD family. B. megaterium
22
PYR2910 P2C decarboxylase, A. niger Fdc1 (XP_001390534),
UbiX
Figure 4. X-band CW EPR spectra of PA0254 . EPR spectra of (A)
2
3
Saccharomyces cerevisiae Fdc1 (NP_010828),
E. coli UbiD
2+
UbiX
Mn in aqueous buffer solution and (B) PA0254 , both recorded
at 20 K using 0.5 mW microwave power and a modulation amplitude
spectrum recorded at 20 K
using 10 μW microwave power and 1 G modulation amplitude.
24
(
ACT45527), Synechocystis sp. PCC 6803 4-hydroxy-3-solanesyl-
benzoate decarboxylase (NP_440197), P. aeruginosa PA0254/
HudA (WP_003115722), Streptomyces griseochromogenes TtnD
25
UbiX
of 7 G. (C) g = 2 region of the PA0254
1
6
2
6
(
(
ABV91289),
Streptomyces himastatinicus SmdK
27 20
WP_009713959), C. basilensis HmfF (ADE20406), Bacillus
28
UbiD clades containing the fungal Fdc1 enzymes and
Streptomyces decarboxylase genes. The latter are involved in
secondary metabolite biosynthesis, and both enzymes are
typically associated with acrylic acid-type substrates (Figure 2).
Despite the distinct heteroaromatic substrate specificity, the B.
megaterium PYR2910 P2C decarboxylase was found to possess
0% identity with Aspergillus niger Fdc1, as opposed to 31%
identity with the Cupriavidus basilensis HmfF furan-2,5-
The nearest homologue
that has been studied in detail is PA0254 from P. aeruginosa
PAO1 (44% identity). Indeed, the first UbiD crystal
structure to be reported was that of P. aeruginosa PA0254.
However, the precise function or activity of PA0254 (also
known as HudA) was unknown, although it has been
implicated as a virulence attenuation factor.
subtilis VdcC (WP_003246683), Thauera aromatica phenylphos-
phate carboxylase alpha (CAC12690) and beta (CAC12688)
2
9
30
subunits, Clostridium BF benzene carboxylase (ADJ94002),
31
Klebsiella pneumoniae AroY (WP_004224180), Lactobacillus planta-
rum 3,4-dihydroxybenzoate decarboxylase LpdC (WP_003644796),
Aromatoleum aromaticum phthalyl-CoA decarboxylase
WP_011255033), Syntrophorhabdus aromaticivorans isophthalyl-
CoA decarboxylase (WP_028893368), Streptomyces tubercidicus
32
33
(
4
34
35
36
TubF (AVW83014), and Aeropyrum pernix (APE_2078). Differ-
ent branches can be grouped by substrate specificity. Structures of
substrates are shown with the leaving group (carboxylate for
decarboxylases, hydrogen for carboxylases) in red. Colored segments
indicate subfamilies for which crystal structures are available.
Sequences of UbiD homologues were aligned using T-coffee, and
Trees were generated using Geneious Tree builder using the
neighbor-joining method.
13,20
dicarboxylic acid decarboxylase.
1
6
16
2
1
As a
2
868
https://dx.doi.org/10.1021/acscatal.0c05042
ACS Catal. 2021, 11, 2865−2878

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 5. Pyrrole-2-carboxylate (P2C) decarboxylase activity of PA0254. (A) UV−vis spectra of P2C following addition of PA0254, revealing a
peak corresponding to pyrrole production (210.5 nm) appears over time. (B) HPLC analysis of P2C incubated with PA0254 confirms formation of
a product with the same retention time as pyrrole. (C) Michaelis−Menten kinetics of P2C decarboxylation.
Figure 6. PA0254 activity is light sensitive. PA0254 activity under
aerobic conditions in the light (red) and in the dark (blue). Assays
were performed against 1 mM P2C. Error bars represent SEM, n = 3.
1
1
Figure 7. H NMR analysis of pyrrole H/D exchange. H NMR
consequence, the structure obtained was that of the apo-
enzyme, lacking the prFMN cofactor that had yet to be
identified at the time. Furthermore, the B. megaterium
spectra of pyrrole in D O when incubated with (reaction) and
2
UbiX
without (ctrl) PA0254
.
2
PYR2910 enzyme, PA0254, and the fungal Fdc1 form
8
,16
dimers,
(
in contrast to the hexameric (hetero)aromatic
PA0254 Is a Pyrrole-2-carboxylate Decarboxylase. We
expressed His-tagged PA0254 with and without UbiX in E. coli.
A comparison of the UV−vis spectra of the purified proteins
reveals UbiX coexpression has a drastic effect on the PA0254
12
13
14
de)carboxylases UbiD, HmfF , and AroY. This confirms
that a quaternary structure or position within the UbiD family
tree does not always correlate with substrate specificity.
2
869
https://dx.doi.org/10.1021/acscatal.0c05042
ACS Catal. 2021, 11, 2865−2878

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 8. PA0254UbiX catalyzed carboxylation of pyrrole to P2C. (A) Schematic of pyrrole carboxylation. (B) UV−vis spectrum of pyrrole following
UbiX
UbiX
addition of PA0254
in the presence of KHCO . Spectra recorded at 1 min intervals following addition of PA0254 . Peak at 255 nm increases
3
in intensity over time corresponding to P2C production. (C) HPLC analysis of pyrrole incubated with PA0254 in the presence of KHCO and/or
3
pressurized CO , confirming that the product has the same retention time as P2C. (D) Relative conversions of pyrrole to P2C under the conditions
2
shown in C. Highest proportion of pyrrole converted to P2C was 55% in the presence of KHCO and 1.5 MPa pressurized CO .
3
2
app
and kcatapp values of 4.3 (±0.5)
properties (Figure 3). In the presence of UbiX (denoted as
PA0254 ), spectral features associated with prFMN are
decarboxylation indicated K
m
UbiX
−1
mM and 35.8 (±0.8) s , respectively (Figure 5C), values
observed, while in the absence of UbiX coexpression, a more
flavin-like spectrum is observed. For the PA0254UbiX sample,
the presence of a minor 550 nm feature is indicative of the
similar to those previously reported for the B. megaterium
homologue. These are reported as apparent values in view of
the presence of a minor population of inactive prFMN radical
species complicating accurate quantification of active enzyme
concentration. However, the relative increase in the prFMN
2
radical
8
presence of the prFMN
semiquinone species. Addition of
UbiX
NaBH CN to PA0254
resulted in the development of a
3
visible purple color under aerobic conditions with a
radical signal following NaBH CN reduction under aerobic
3
corresponding increase in the A550 peak. This is consistent
conditions suggests the signal in the as-isolated samples
iminium
with conversion of the major prFMN
species to the
accounts for less than 10% of the total population.
radical
8
UbiX
prFMN
as previously reported for A. niger Fdc1.
PA0254
Activity Is Light Sensitive. Previous reports
The prFMNradical so formed is detectable in the X band CW
EPR spectrum (Figure 4) at the center of the six line pattern
indicated that the activity of the B. megaterium enzyme was
oxygen sensitive, with the addition of reducing reagents
2
+
37
arising from the m = ±1/2 manifold of the S = 5/2 Mn ion,
required to stabilize the enzyme. Fdc1 has also been reported
s
8
the six lines arising from hyperfine interaction with the I = 5/2
to lose activity over time, in this case the result of light-
5
5
iminium 38
Mn nucleus. Comparison of Figure 4B with spectrum Figure
induced isomerization and inactivation of prFMN
Similar to Fdc1, PA0254
.
2
+
UbiX
4
A demonstrates that this is protein-bound, not free, Mn , and
was found to lose activity with a
Figure 4C presents the radical signal recorded under low
power and modulation amplitude to show that it has the same
form as the prFMNradical: Mn -coupled signal of A. niger Fdc1
half-life of ∼140 min when incubated on ice in a transparent
tube. By contrast, enzyme activity was found to be stable when
2+
kept in the dark (Figure 6).
UbiX
(
8) and E. coli UbiD (12).
PA0254
Catalyzes Pyrrole H/D Exchange. Previous
As previously reported by Jacewicz et al., no decarboxylation
work has shown that UbiD decarboxylases are capable of
activity could be detected with the Fdc1 substrate cinnamic
acid. However, when PA0254UbiX was added to pyrrole-2-
carboxylate, pyrrole was readily formed as confirmed by UV−
vis (Figure 5A) and HPLC analysis (Figure 5B). In contrast,
PA0254 (i.e., expressed in the absence of UbiX) possessed
little or no activity, confirming prFMN is required for P2C
decarboxylation. The Michaelis−Menten kinetics for P2C
catalyzing deuterium exchange on the decarboxylation reaction
39
1
products. H NMR showed that incubation of pyrrole with
UbiX
PA0254
in D O resulted in depletion of the resonance peak
2
at 6.9 ppm, consistent with exchange of both protons in
positions 2 and 5 (denoted H ) with deuterons (Figure 7).
a
UbiX
PA0254
Catalyzes Pyrrole Carboxylation at Ele-
vated [CO ]. B. megaterium PYR2910 was previously reported
2
2
870
https://dx.doi.org/10.1021/acscatal.0c05042
ACS Catal. 2021, 11, 2865−2878

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 9. PA0254UbiX substrate scope. Activity of PA0254UbiX with heteroaromatic substrates including (A) pyrrole-2-carboxylate (P2C), (B) furan-
-carboxylate (F2C), (C) thiophene-2-carboxylate (T2C), and (D) indole-3-carboxylate (I3C). Ten millimolar product standards are shown in
2
blue, while 10 mM substrate with and without enzyme are shown in green and red, respectively. Panels on the right show a zoom in of the adjacent
chromatograms. See Figure 15C for a comparison of substrate depletion levels following addition of and incubation with PA0254UbiX for a range of
heteroaromatic acids.
to catalyze the carboxylation of pyrrole in the presence of
elevated concentrations of CO , either through CO at high
incubated overnight with 50 mM pyrrole and 1 M potassium
bicarbonate and/or under pressurized CO (1.5 MPa) revealed
2
2
2
pressure or via addition of high concentrations of bicarbonate.
a peak with a retention time of 2.5 min that comigrates with a
P2C standard (Figure 8C). The highest proportion of pyrrole
converted to P2C was in the presence of both 1 M KHCO3
UbiX
To investigate the ability of PA0254
to catalyze the
carboxylation of pyrrole, the enzyme was incubated with 25
mM pyrrole and 0.5 M potassium bicarbonate. UV−vis spectra
and 1.5 MPa CO (Figure 8D), whereas no P2C could be
2
UbiX
recorded at 1 min intervals after addition of PA0254
detected in the absence of additional bicarbonate/CO or
2
revealed that a peak centered at 255 nm appeared and
increased over time, consistent with the production of P2C
with a kobs of 4.6 s− (Figure 8). HPLC analysis of reactions
enzyme.
UbiX
PA0254
Substrate Specificity Is Restricted.
1
UbiX
PA0254
activity was screened with various acids in order
2
871
https://dx.doi.org/10.1021/acscatal.0c05042
ACS Catal. 2021, 11, 2865−2878

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 10. Activity of PA0254UbiX with imidazole substrates. HPLC chromatograms of (A) imidazole-2-carboxylate (Im2C) and (B) imidazole-4-
carboxylate (Im4C). Ten millimolar product standards are shown in blue, while 10 mM substrate with and without enzyme are shown in green and
red, respectively. Lower panels show a zoom in to the above chromatograms; 0% of Im2C and 1.3% of Im4C is converted under the conditions
tested. (C and D) H NMR analysis of Im2C (C) and Im4C (D) with (reaction) and without (ctrl) PA0254UbiX.
1
2+
+
to determine the substrate range. Evidence of decarboxylase
activity could be found with 3-methylpyrrole-2-carboxylate,
indole-3-carboxylate, and furan-2-carboxylate with very low
activity in the case of thiophene-2-carboxylate (Figure 9). In
contrast, no evidence for the decarboxylation of pyrrole-3-
that PA0254UbiX requires both Mn and K for activity (Figure
12). The overall conformation and position of the prFMN
binding domain is highly similar in the four monomers present
in the AU and resembles the closed conformation previously
observed for the A. niger Fdc1. Comparison with the previously
reported apo-PA0254 structures reveals that considerable
reorientation occurs in various loop regions associated with
carboxylate, indole-2-carboxylate, or benzoic acid was observed
UbiX
under the conditions tested. To test whether PA0254
could
2
+
also decarboxylate imidazole-based compounds, we incubated
the enzyme with imidazole-2-carboxylate (Im2C) and
imidazole-4-carboxylate (Im4C). A small quantity of imidazole
was produced in an enzyme-dependent manner from Im4C,
while enzyme-dependent decarboxylation of Im2C was not
observed (Figure 10). This was confirmed by monitoring the
Mn and/or prFMN binding.
Furthermore, the relative position of the prFMN binding
domain with respect to the C-terminal dimerization domain is
distinct in the apo-structures, adopting a more open
conformation. This suggests that the open-to-closed transition
is affected by prFMN and/or substrate binding. In the case of
the related A. niger Fdc1, cofactor binding has indeed been
demonstrated to affect the protein overall conformation using
1
UbiX
reaction using H NMR. In the presence of PA0254 , the
1
substrate-derived H peaks for both Im2C and Im4C
40
broadened, possibly indicating binding to the enzyme. Two
new peaks corresponding to imidazole formed only with Im4C
mass spectrometry.
UbiX
PA0254
Crystal Structure Yields an Imidazole
UbiX
(
Figure 10C and 10D).
Adduct. The PA0254
active site is readily identified by the
UbiX
PA0254
Crystal Structure Determination.
was screened against 480 crystallization conditions
presence of the prFMN cofactor. In each of the four
monomers, additional electron density was observed located
above the prFMN cofactor (Figure 13). The density was
UbiX
PA0254
and found to readily crystallize in a range of conditions.
Unfortunately, several crystal forms suffered from various
twinning pathologies, hindering structure determination.
Crystals obtained in the presence of 100 mM imidazole
modeled as an imidazole derived from the crystal mother
UbiX
liquor. In two of the PA0254
monomers, the imidazole
density was continuous with the prFMN density, suggesting a
covalent bond between the imidazole C2 and the C1′ of the
prenyl-derived prFMN ring. The electron density in the other
two monomers was noncontinuous between the imidazole and
prFMN and thus modeled as a noncovalent, stacking
interaction between imidazole and the prFMN aromatic
plane. In both the covalent and the noncovalent ligand
complexes, the imidazole N1 is within the hydrogen-bonding
distance of the N318 amide side chain. The H-bonding
(
acting as the buffering agent) did occasionally yield
nontwinned crystals which were used for data collection and
refinement. A structure was obtained to 1.65 Å resolution
using the apo-PA0254 structure as a molecular replacemUebniXt
model. The asymmetric unit (AU) contains two PA0254
dimers, with each monomer found to bind a prFMN cofactor
2
+
+
via Mn and K coordination (Figure 11). The identification
of both ions is based on the electron density, the observation
2
872
https://dx.doi.org/10.1021/acscatal.0c05042
ACS Catal. 2021, 11, 2865−2878

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
deviation of approximately 35°) suggests either protonation of
3
C2 (i.e., sp hybridization, nonaromatic) or considerable strain
2
on the aromatic sp -hybridized C2 (i.e., Int 2-like) imposed by
the protein active site. Previous studies on A. niger Fdc1 have
confirmed that the active site is able to constrain prFMN
adducts, thus controlling the internal thermodynamics of the
11
reaction.
Reversible Binding of Imidazole to PA0254UbiX
Following the identification of the imidazole adduct in the
.
UbiX
PA0254
crystal structure, the enzyme was subsequently
purified using histidine as elutant as opposed to imidazole. A
comparison of the UV−vis profiles of the imidazole- and
histidine-eluted enzymes revealed subtle differences with the
imidazole-eluted protein having a slightly lower and more
defined peak at ∼340 nm (Figure 14). Upon incubation of the
histidine-eluted protein with imidazole, a small spectral shift
was observed. Mass spectrometry of the imidazole-bound
species revealed a peak at 593.2094 Da, close to the predicted
mass of a covalently bound imidazole−prFMN adduct
(
593.2125 Da). Both imidazole and histidine purified proteins
were found to have similar activities. The activity of the
histidine-eluted protein was inhibited by ∼50% by addition of
1
0 mM imidazole. Full activity was recovered following
removal of imidazole by desalting, suggesting that reversible
prFMN adduct formation occurs.
Figure 11. Cofactor binding in PA0254UbiX. Overlay of the apo-
UbiX
PA0254 structure (in blue) with the PA0254
holo-enzyme
Mutagenesis of PA0254 Supports a Key Role for
N318. A number of PA0254 variants were generated in order
to establish the proposed role of N318 in substrate recognition.
In the case of HmfF, the corresponding residue is H297, which
was established to be key to ensuring specificity for furan-type
substrates. Hence, we designed a PA0235 N318H variant to
test whether this would affect the substrate specificity and
possibly alter the preference to furan-type substrates.
Surprisingly, the purified N318H variant was found to possess
an intense yellow color consistent with oxidized flavin (Figure
structure (in red/cyan) reveals the distinct position of the
oligomerization domain with respect to the prFMN binding domain.
Domain motion required to change from the apo-PA0254
UbiX
conformation (in blue) to the holo-PA0254
(in red) leads to
closure of the prFMN binding cleft, concomitant with active site
2
+
closure. While the apo-PA0254 structure contains Mg , the holo-
UbiX
2+
+
PA0254
contains Mn and K ions that establish an ionic network
with the prFMN phosphate group.
1
5). Structure determination of the N318H variant revealed
that the introduced histidine side chain partially occupies the
site of the prenyl-derived fourth prFMN ring, preventing
prFMN binding. Hence, we created a series of distinct N318X
variants, restricted to amino acids with a size similar to or
smaller than asparagine. While all variants appeared to bind
cofactor, this unfortunately occurs to varying levels as judged
by the UV−vis spectra, making detailed comparison of the
UbiX
relative activity levels complicated. The PA0254
variants
were screened for decarboxylation activity with P2C and furan-
2-carboxylic acid (F2C). Only the WT and N318D enzymes
were able to achieve complete decarboxylation of the P2C
substrate. Of the variants tested, N318C and N318S also had
considerable levels of activity whereas replacement of N318
with nonpolar residues resulted in very low conversion,
possibly due to low cofactor incorporation. The N318S and
N318C preparations yielded a higher conversion of the F2C
and T2C substrate compared to the WT enzyme under the
conditions tested.
2
+
+
Figure 12. PA0254 activity is Mn and K dependent. Bar chart of
apo-PA0254 P2C decarboxylation activity following reconstitution
+
+
2+
2+
with either Na /K and Mg /Mn a+nd prFMN. Activity was highest
2
+
in the presence of both Mn and K ions.
interaction of N318 with E316 suggests that the N318 amide is
indeed oriented with the oxygen toward the imidazole
nitrogen. Hydrophobic interactions with Y279, W322, F432,
and M434 largely occlude the imidazole from the solvent, with
the exception of the water molecule hydrogen bonding to the
imidazole N3. In the case of the covalent imidazole adduct, the
imidazole C2 is positioned 3.2 Å away from the conserved
E278. The latter has been implicated as the key active site
acid−base catalyst required to either donate or abstract a
proton from the prFMN-bound substrate. At 1.65 Å resolution
it is not possible to directly determine the protonation state of
the imidazole C2 for the covalent adduct. However, the fact
that the C1′ is not in the plane with the imidazole moiety (a
DFT Calculations Reveal Two Int1 Species. In order to
better understand how the P2C, F2C, and T2C substrates
react with the prFMN, density functional theory (DFT)
calculations of prFMN adducts formed with P2C, F2C, T2C,
Im2C, and Im4C were carried out. These calculations were
performed in implicit water rather than the enzyme active site
for simplicity and to facilitate the direct comparison of the role
open
of the adduct moiety. In all cases, stable ring-open Int1
C1′−C2 adducts associated with a nucleophilic or electrophilic
2
873
https://dx.doi.org/10.1021/acscatal.0c05042
ACS Catal. 2021, 11, 2865−2878

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 13. PA0254UbiX active site structure. Active site residues shown in atom-colored sticks (gray carbons) with the prFMNiminium cofactor shown
with yellow carbons. Bound imidazole derived from the crystallization buffer is shown with green carbons. Omit F F map corresponding to
o
c
cofactor and imidazole is contoured at 3 sigma and shown as a cyan mesh. In two monomers a covalent bond is formed between imidazole C2 and
the prFMN C1′ (top view), while the active site of the other monomers lacks electron density in between the imidazole and the prFMN, indicating
a noncovalent complex. Hydrogen bonds are shown in yellow dotted lines, while the key Glu278 imidazole C2 and imidazole C2/prFMN C1′
interactions are shown in red.
addition process were observed. This species has been
observed in previous DFT active site “cluster” calculations
where decarboxylation appears to occur from this rather than
ylase that acts predominantly on cinnamic acid-type sub-
8
,42,43
strates.
In this case, sufficient evidence has accumulated
that supports a reversible 1,3-dipolar cycloaddition mechanism
closed
11,15
+
11
the ring-closed Int1 species (Int1
) shown in Figure 1.
underpinning the (de)carboxylation reaction. Chemical
With the notable exception of the zwitteronic Im2C−H
species (with protons on both N1 and N3), the Int1closed
species was found to be stable for all adducts and lower in
energy compared to the corresponding Int1open adduct (Table
S1). A natural charge analysis shows that in all cases, except for
precedent exists for the reaction of cinnamic acid-type
dipolarophiles with azomethine ylide species, and the proposed
mechanism also provides an explanation for the need of the
iminium
elaborate FMN to prFMN
transformation.
However, the substrate scope of the wider UbiD family
extends far beyond cinnamic acid substrates, including both
+
Im2C−H , there is partial electron transfer from the substrate
open
closed
6
to the prFMN to form Int1 . The Int1
adduct is
heteroaromatic and aromatic acids. It is clear that the latter
subsequently formed by electron transfer back from the
prFMN to the substrate moiety of the adduct (Table S2,
substrates have inherently different reactivity and impose
distinct conformational and energetic challenges for the
enzyme. In the case of A. niger ferulic acid decarboxylase,
variants have been developed that accept (hetero)aromatic
Supporting Information). This is consistent with the formation
open
of Int1
occurring by nucleophilic attack from the substrate,
+
except for nonsubstrate Im2C−H which acts as an electro-
phile. Subsequently, the substrate Int1
15
acids with low reactivity observed for 2-naphthoic acid. In
open
adducts formed by
this case, an electrophilic aromatic substitution process has
nucleophilic attack appear to have Wheland intermediate
character.
been proposed with stabilization of the charge on the Wheland
iminium
intermediate through stacking with the prFMN
. In
contrast, the decarboxylation of 3,4-dihydroxybenzoic acid
substrates by AroY is proposed to occur via a quinoid
intermediate formed concomitant with prFMN C1′−substrate
DISCUSSION
■
The unusual metamorphosis of flavin to prFMNiminium alters
the fundamental character of this cofactor. In contrast to the
C4a/N5 focused reactivity of flavin, the N5−C6 prenylation
and subsequent oxidative maturation of prFMN
14
C alpha bond formation. Finally, reversible decarboxylation
of furan dicarboxylic acid by HmfF has been proposed to occur
through either a cycloaddition or an electrophilic aromatic
iminium
lead to a
13
stabilized azomethine ylide species with a reactive C4a/N5/
substitution process.
8
,41
iminium
C1′ center.
Assuming prFMN
underpins catalysis in
In the case of PA0254, the structure of the covalent
iminium
all UbiD enzymes, certain general principles are likely to apply
across this ubiquitous microbial enzyme family. Arguably, the
best-understood enzyme is the A. niger ferulic acid decarbox-
prFMN
C1′−imidazole C2 adduct provides further
insight into the general reaction of UbiD enzymes with
heteroaromatic compounds. Crucially, the imidazole-2-carbox-
2
874
https://dx.doi.org/10.1021/acscatal.0c05042
ACS Catal. 2021, 11, 2865−2878

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 14. Imidazole binding to PA0254UbiX. (A) UV−vis spectra of PA0254UbiX eluted from IMAC using imidazole (blue) or histidine (red).
Spectrum of the histidine-eluted protein following addition of 20 mM imidazole and desalting is also shown (green). (B) Mass spectrum of low
UbiX
molecular weight species within PA0254
eluted with imidazole, indicating a species with a mass of 593.2094 Da, close in mass to the predicted
UbiX
covalent imidazole−prFMN adduct. (C) Activity of PA0254
possess similar activities. Addition of 10 mM imidazole results in ∼50% inhibition; however, activity is recovered upon removal of excess imidazole
error bars represent SEM, n = 3).
against P2C following various treatments. Imidazole- and histidine-eluted proteins
(
Figure 15. PA0254UbiX N318X characterization. (A) UV−vis spectra of PA0254 variants normalized on the A280 peak. (Inset) Zoom into cofactor-
UbiX
related features in the 300−800 nm region. (B) Overlay of the WT PA0254
crystal structure (with gray carbons) with the N318H mutant (in
blue carbons) reveals bound FMN (yellow carbons). Omit F F map corresponding to FMN is contoured at 3 sigma and shown as a cyan mesh.
o
c
(
C) Comparison of substrate depletion levels following addition of and incubation with WT, N318C, and N318S variants for a range of
heteroaromatic acids. Wavelength used for quantification shown in brackets. Error bars represent SEM, n = 3.
ylic acid is not a substrate for the enzyme, while imidazole acts
as a reversible inhibitor. The chemical reactivity of imidazole
for electrophilic aromatic substitution is substantially lower
than that of the corresponding pyrrole and only occurs on C4/
2
875
https://dx.doi.org/10.1021/acscatal.0c05042
ACS Catal. 2021, 11, 2865−2878

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 16. Proposed mechanism for PA0254/HudA. prFMNiminium electrophilic aromatic substitution reaction with pyrrole underpins reversible
decarboxylation, while nucleophilic addition to the imidazole C2 position leads to reversible inhibition.
C5 positions. In contrast, the imidazole C2 position is electron
deficient and undergoes nucleophilic aromatic substitution
when a suitable leaving group is present. This suggests that
reversible bond formation with imidazole C2 might occur
through nucleophilic attack of C1′ concomitant with
protonation at N3. Crucially, this prFMN adduct species
aromaticity of the benzene ring. However, UbiD enzymes have
been implicated in microbial anaerobic benzene degradation
where carboxylation is proposed to activate the substrate for
45
further degradation. We have not been able to establish
UbiX
whether benzoic acid can bind to PA0254 , while variants
aimed at creating a more hydrophobic active site (i.e., N318A/
V/L) did not readily bind prFMN. It thus remains possible
that a UbiD enzyme with an active site optimized for benzene/
benzoic acid binding might be able to catalyze electrophilic
aromatic substitution at rates sufficient to support the relatively
slow microbial growth seen during anaerobic benzene
degradation. Furthermore, the domain dynamics indirectly
observed here, but not invoked for either Fdc1 or PA0254
reaction, might couple to the reaction coordinate in the case of
more challenging transformations such as benzene/napthalene
(
labeled Inhib in Figure 16) would not support H/D exchange
or (de)carboxylation at the C2 position, in line with our
observations in solution.
In contrast, reaction with pyrrole/furan/thiophene com-
pounds is likely to occur through electrophilic aromatic
substitution at the C2 position via a Wheland-type
intermediate Int1open/Int3open (Figure 16). While DFT
closed
calculations indicate an Int1
species might occur, it is
unclear what role it plays in catalysis. It is however interesting
29,45,46
closed
or phenylphosphate carboxylation.
to note the Int1
species appears inaccessible to the
While the exact biological role of PA0254/HudA as a
pyrrole-2-decarboxylase is yet to be established, the recent
observation that P2C eliminates the expression of quorum
sensing cascade and pathogenic factors of P. aeruginosa PAO1
zwitterionic imidazole-2-carboxylic acid. The Int1 and Int3
intermediates provide access to the central Int2 species via,
respectively, (de)carboxylation and (de)protonation. It is
possible that additional through-space electronic interactions
47
iminium
on both phenotypic and genotypic levels suggests that it
could be involved in P2C detoxification. Previous identification
of PA0254/HudA as a virulence attenuation factor, on the
with the prFMN
stabilize the charge on the Wheland
4
4
intermediate and assist with retaining the strained config-
uration of Int2. The latter configuration is required within the
context of the closed active site as highlighted by the
2
1
other hand, might indicate that the product of the
decarboxylation reaction, pyrrole, is responsible for the
observed effects.
UbiX
PA0254 −imidazole complex. A configuration whereby the
C1′ prFMN substituent is positioned in the plane of the
substrate aromatic ring would require active site reorganiza-
tion, similar to the domain motion observed when comparing
the apo- and holo-forms of PA0254.
ASSOCIATED CONTENT
sı Supporting Information
■
*
However, to ensure rapid turnover, highly stable covalent
intermediates should be avoided, and we propose that the
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acscatal.0c05042.
HF energy (hartrees) computed for the Int1closed
aromatic group in the Int2 adduct remains parallel rather than
iminium
perpendicular to the prFMN
plane. The trend in yields
adducts; natural charges computed for the substrate
obtained with pyrrole-, furan-, and thiophene-2-carboxylic acid
compounds mirrors the respective reactivity toward electro-
philic aromatic substitution in solution. No activity with
pyrrole-3-carboxylic acid could be detected, but the related
indole-3-carboxylic acid readily yielded indole. This again
mirrors the trends for electrophilic aromatic substitution
reactivity, which is preferred for pyrrole at the 2 position
while indole occurs at the 3 position. Unfortunately, no
reaction could be observed with benzoic acid, a substrate that
arguably presents the most formidable barrier due to the high
open
closed
moiety; alignment of the Int1
and Int1
; DFT
models used in this study (PDF)
AUTHOR INFORMATION
■
Corresponding Author
David Leys − Manchester Institute of Biotechnology, University
of Manchester, Manchester M1 7DN, United Kingdom;
orcid.org/0000-0003-4845-8443; Phone: 0044 161 306
51 50; Email: david.leys@manchester.ac.uk
2
876
https://dx.doi.org/10.1021/acscatal.0c05042
ACS Catal. 2021, 11, 2865−2878

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Authors
(4) Yoshida, T.; Fujita, K.; Nagasawa, T. Novel reversible indole-3-
carboxylate decarboxylase catalyzing nonoxidative decarboxylation.
Biosci., Biotechnol., Biochem. 2002, 66, 2388−94.
Karl A. P. Payne − Manchester Institute of Biotechnology,
University of Manchester, Manchester M1 7DN, United
Kingdom; orcid.org/0000-0002-6331-6374
Stephen A. Marshall − Manchester Institute of Biotechnology,
University of Manchester, Manchester M1 7DN, United
Kingdom
Karl Fisher − Manchester Institute of Biotechnology, University
of Manchester, Manchester M1 7DN, United Kingdom
Stephen E. J. Rigby − Manchester Institute of Biotechnology,
University of Manchester, Manchester M1 7DN, United
Kingdom
Matthew J. Cliff − Manchester Institute of Biotechnology,
University of Manchester, Manchester M1 7DN, United
Kingdom
Reynard Spiess − Manchester Institute of Biotechnology,
University of Manchester, Manchester M1 7DN, United
Kingdom
Diego M. Cannas − Department of Chemistry, University of
Manchester, Manchester M13 9PL, United Kingdom
Igor Larrosa − Department of Chemistry, University of
Manchester, Manchester M13 9PL, United Kingdom;
orcid.org/0000-0002-5391-7424
Sam Hay − Manchester Institute of Biotechnology, University
of Manchester, Manchester M1 7DN, United Kingdom;
orcid.org/0000-0003-3274-0938
(5) Abu Laban, N.; Selesi, D.; Rattei, T.; Tischler, P.; Meckenstock,
R. U. Identification of enzymes involved in anaerobic benzene
degradation by a strictly anaerobic iron-reducing enrichment culture.
Environ. Microbiol. 2010, 12 (10), 2783−2796.
(6) Marshall, S. A.; Payne, K. A. P.; Leys, D. The UbiX-UbiD
system: The biosynthesis and use of prenylated flavin (prFMN). Arch.
Biochem. Biophys. 2017, 632, 209−221.
(7) Mukai, N.; Masaki, K.; Fujii, T.; Kawamukai, M.; Iefuji, H. PAD1
and FDC1 are essential for the decarboxylation of phenylacrylic acids
in Saccharomyces cerevisiae. J. Biosci. Bioeng. 2010, 109, 564−9.
(8) Payne, K. A.; White, M. D.; Fisher, K.; Khara, B.; Bailey, S. S.;
Parker, D.; Rattray, N. J.; Trivedi, D. K.; Goodacre, R.; Beveridge, R.;
Barran, P.; Rigby, S. E.; Scrutton, N. S.; Hay, S.; Leys, D. New
cofactor supports alpha, beta-unsaturated acid decarboxylation via 1,3-
dipolar cycloaddition. Nature 2015, 522, 497−501.
(9) White, M. D.; Payne, K. A.; Fisher, K.; Marshall, S. A.; Parker,
D.; Rattray, N. J.; Trivedi, D. K.; Goodacre, R.; Rigby, S. E.; Scrutton,
N. S.; Hay, S.; Leys, D. UbiX is a flavin prenyltransferase required for
bacterial ubiquinone biosynthesis. Nature 2015, 522, 502−6.
(10) Marshall, S. A.; Payne, K. A. P.; Fisher, K.; White, M. D.; Ni
Cheallaigh, A.; Balaikaite, A.; Rigby, S. E. J.; Leys, D. The UbiX flavin
prenyltransferase reaction mechanism resembles class I terpene
cyclase chemistry. Nat. Commun. 2019, 10, 2357.
(11) Bailey, S. S.; Payne, K. A. P.; Saaret, A.; Marshall, S. A.;
Gostimskaya, I.; Kosov, I.; Fisher, K.; Hay, S.; Leys, D. Enzymatic
control of cycloadduct conformation ensures reversible 1,3-dipolar
cycloaddition in a prFMN-dependent decarboxylase. Nat. Chem.
2019, 11, 1049−1057.
Complete contact information is available at:
https://pubs.acs.org/10.1021/acscatal.0c05042
(12) Marshall, S. A.; Fisher, K.; Ni Cheallaigh, A.; White, M. D.;
Author Contributions
Payne, K. A.; Parker, D. A.; Rigby, S. E.; Leys, D. Oxidative
Maturation and Structural Characterization of Prenylated FMN
Binding by UbiD, a Decarboxylase Involved in Bacterial Ubiquinone
Biosynthesis. J. Biol. Chem. 2017, 292, 4623−4637.
D.L. conceived and coordinated the study. K.A.P.P. and S.A.M.
carried out experiments. M.J.C ran NMR experiments. R.S.
performed mass spectrometry. K.F. and S.E.J.R. ran EPR
experiments. D.M.C. and I.L. assisted with carboxylation
experiments. D.L. wrote the paper. All authors reviewed the
results and approved the final version of the manuscript.
(13) Payne, K. A. P.; Marshall, S. A.; Fisher, K.; Cliff, M. J.; Cannas,
D. M.; Yan, C.; Heyes, D. J.; Parker, D. A.; Larrosa, I.; Leys, D.
Enzymatic Carboxylation of 2-Furoic Acid Yields 2,5-Furandicarbox-
ylic Acid (FDCA). ACS Catal. 2019, 9, 2854−2865.
Notes
(14) Payer, S. E.; Marshall, S. A.; Barland, N.; Sheng, X.; Reiter, T.;
Dordic, A.; Steinkellner, G.; Wuensch, C.; Kaltwasser, S.; Fisher, K.;
Rigby, S. E. J.; Macheroux, P.; Vonck, J.; Gruber, K.; Faber, K.; Himo,
F.; Leys, D.; Pavkov-Keller, T.; Glueck, S. M. Regioselective para-
Carboxylation of Catechols with a Prenylated Flavin Dependent
Decarboxylase. Angew. Chem., Int. Ed. 2017, 56, 13893−13897.
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
This work was supported by the European Research Council
ERC) grant pre-FAB ADG_695013 and BBSRC grant BB/
■
(
(15) Aleku, G. A.; Saaret, A.; Bradshaw-Allen, R. T.; Derrington, S.
P000622/1. The atomic coordinates and structures factors
codes 7ABN and 7ABO) have been deposited to the Protein
R.; Titchiner, G. R.; Gostimskaya, I.; Gahloth, D.; Parker, D. A.; Hay,
S.; Leys, D. Enzymatic C-H activation of aromatic compounds
through CO2 fixation. Nat. Chem. Biol. 2020, 16, 1255−1260.
(
Data Bank (http://www.pdb.org). We thank Diamond Light
Source for access (proposal number MX12788) that
contributed to the results presented here. We also thank
Prof. Toyokazu Yoshida for providing us with the sequence of
the B. megaterium PYR2910 pyrrole-2-carboxylate decarbox-
ylase. D. L. is a Royal Society Wolfson Research Merit Award
holder.
(16) Jacewicz, A.; Izumi, A.; Brunner, K.; Schnell, R.; Schneider, G.
Structural insights into the UbiD protein family from the crystal
structure of PA0254 from Pseudomonas aeruginosa. PLoS One 2013, 8,
No. e63161.
(17) Winn, M. D.; Ballard, C. C.; Cowtan, K. D.; Dodson, E. J.;
Emsley, P.; Evans, P. R.; Keegan, R. M.; Krissinel, E. B.; Leslie, A. G.;
McCoy, A.; McNicholas, S. J.; Murshudov, G. N.; Pannu, N. S.;
Potterton, E. A.; Powell, H. R.; Read, R. J.; Vagin, A.; Wilson, K. S.
Overview of the CCP4 suite and current developments. Acta
Crystallogr., Sect. D: Biol. Crystallogr. 2011, 67, 235−42.
(18) Kabsch, W. XDS. Acta Crystallogr., Sect. D: Biol. Crystallogr.
2010, 66, 125−32.
(19) Grimme, S.; Ehrlich, S.; Goerigk, L. Effect of the damping
function in dispersion corrected density functional theory. J. Comput.
Chem. 2011, 32, 1456−65.
(20) Koopman, F.; Wierckx, N.; de Winde, J. H.; Ruijssenaars, H. J.
Identification and characterization of the furfural and 5-
(hydroxymethyl)furfural degradation pathways of Cupriavidus basi-
lensis HMF14. Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 4919−24.
REFERENCES
■
(
1) Lupa, B.; Lyon, D.; Gibbs, M. D.; Reeves, R. A.; Wiegel, J.
Distribution of genes encoding the microbial non-oxidative reversible
hydroxyarylic acid decarboxylases/phenol carboxylases. Genomics
2
(
005, 86, 342−51.
2) Omura, H.; Wieser, M.; Nagasawa, T. Pyrrole-2-carboxylate
decarboxylase from Bacillus megaterium PYR2910, an organic-acid-
requiring enzyme. Eur. J. Biochem. 1998, 253, 480−4.
(3) Ebenau-Jehle, C.; Mergelsberg, M.; Fischer, S.; Bruls, T.;
Jehmlich, N.; von Bergen, M.; Boll, M. An unusual strategy for the
anoxic biodegradation of phthalate. ISME J. 2017, 11, 224−236.
2
877
https://dx.doi.org/10.1021/acscatal.0c05042
ACS Catal. 2021, 11, 2865−2878

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
(
21) Kim, S. H.; Park, S. Y.; Heo, Y. J.; Cho, Y. H. Drosophila
Ferulic Acid Decarboxylase Probed by Isotope Effects and Linear
Free-Energy Relationships. Biochemistry 2016, 55, 2857−63.
(40) Beveridge, R.; Migas, L. G.; Payne, K. A. P.; Scrutton, N. S.;
Leys, D.; Barran, P. E. Mass spectrometry locates local and allosteric
conformational changes that occur on cofactor binding. Nat. Commun.
2016, 7, 12163.
(41) Leys, D. Flavin metamorphosis: cofactor transformation
through prenylation. Curr. Opin. Chem. Biol. 2018, 47, 117−125.
(42) Stratford, M.; Plumridge, A.; Pleasants, M. W.; Novodvorska,
M.; Baker-Glenn, C. A.; Pattenden, G.; Archer, D. B. Mapping the
structural requirements of inducers and substrates for decarboxylation
of weak acid preservatives by the food spoilage mould Aspergillus niger.
Int. J. Food Microbiol. 2012, 157, 375−83.
melanogaster-based screening for multihost virulence factors of
Pseudomonas aeruginosa PA14 and identification of a virulence-
attenuating factor, HudA. Infect. Immun. 2008, 76, 4152−62.
(
22) Plumridge, A.; Melin, P.; Stratford, M.; Novodvorska, M.;
Shunburne, L.; Dyer, P. S.; Roubos, J. A.; Menke, H.; Stark, J.; Stam,
H.; Archer, D. B. The decarboxylation of the weak-acid preservative,
sorbic acid, is encoded by linked genes in Aspergillus spp. Fungal
Genet. Biol. 2010, 47, 683−92.
(
23) Mukai, N.; Masaki, K.; Fujii, T.; Kawamukai, M.; Iefuji, H.
PAD1 and FDC1 are essential for the decarboxylation of phenylacrylic
acids in Saccharomyces cerevisiae. J. Biosc. Bioeng. 2010, 109, 564−569.
(
24) Zhang, H.; Javor, G. T. Identification of the ubiD Gene on the
Escherichia coli Chromosome. J. Bacteriol. 2000, 182, 6243−6246.
(43) Aleku, G. A.; Prause, C.; Bradshaw-Allen, R. T.; Plasch, K.;
Glueck, S. M.; Bailey, S. S.; Payne, K. A. P.; Parker, D. A.; Faber, K.;
Leys, D. Terminal Alkenes from Acrylic Acid Derivatives via Non-
Oxidative Enzymatic Decarboxylation by Ferulic Acid Decarboxylases.
ChemCatChem 2018, 10, 3736−3745.
(25) Pfaff, C.; Glindemann, N.; Gruber, J.; Frentzen, M.; Sadre, R.
Chorismate pyruvate-lyase and 4-hydroxy-3-solanesylbenzoate decar-
boxylase are required for plastoquinone biosynthesis in the
cyanobacterium Synechocystis sp. PCC6803. J. Biol. Chem. 2014,
(44) Murphy, K. E.; Bocanegra, J. L.; Liu, X.; Chau, H. K.; Lee, P.
2
(
89, 2675−2686.
C.; Li, J.; Schneebeli, S. T. Precise through-space control of an abiotic
electrophilic aromatic substitution reaction. Nat. Commun. 2017, 8,
26) Luo, Y.; Li, W.; Ju, J.; Yuan, Q.; Peters, N. R.; Hoffmann, F. M.;
Huang, S.-X.; Bugni, T. S.; Rajski, S.; Osada, H.; Shen, B. Functional
Characterization of TtnD and TtnF Unveiling New Insights into
Tautomycetin Biosynthesis. J. Am. Chem. Soc. 2010, 132, 6663.
1
(
4840.
45) Luo, F.; Gitiafroz, R.; Devine, C. E.; Gong, Y.; Hug, L. A.;
Raskin, L.; Edwards, E. A. Metatranscriptome of an anaerobic
benzene-degrading, nitrate-reducing enrichment culture reveals
involvement of carboxylation in benzene ring activation. Appl.
Environ. Microbiol. 2014, 80, 4095−107.
(27) Wang, B.; Song, Y.; Luo, M.; Chen, Q.; Ma, J.; Huang, H.; Ju, J.
Biosynthesis of 9-methylstreptimidone involves a new decarboxylative
step for polyketide terminal diene formation. Org. Lett. 2013, 15,
1
(
278−1281.
(46) Mouttaki, H.; Johannes, J.; Meckenstock, R. U. Identification of
28) Lupa, B.; Lyon, D.; Shaw, L. N.; Sieprawska-Lupa, M.; Wiegel,
naphthalene carboxylase as a prototype for the anaerobic activation of
non-substituted aromatic hydrocarbons. Environ. Microbiol. 2012, 14,
J. Properties of the reversible nonoxidative vanillate/4-hydroxyben-
zoate decarboxylase from Bacillus subtilis. Can. J. Microbiol. 2008, 54,
2
(
770−4.
7
(
5−81.
47) Hassan, R.; Shaaban, M. I.; Abdel Bar, F. M.; El-Mahdy, A. M.;
29) Schuhle, K.; Fuchs, G. Phenylphosphate carboxylase: a new C-
Shokralla, S. Quorum Sensing Inhibiting Activity of Streptomyces
coelicoflavus Isolated from Soil. Front. Microbiol. 2016, 7, 659.
C lyase involved in anaerobic phenol metabolism in Thauera
aromatica. J. Bacteriol. 2004, 186, 4556−67.
(30) Abu Laban, N.; Selesi, D.; Rattei, T.; Tischler, P.; Meckenstock,
R. U. Identification of enzymes involved in anaerobic benzene
degradation by a strictly anaerobic iron-reducing enrichment culture.
Environ. Microbiol. 2010, 12, 2783−2796.
(31) Niu, W.; Draths, K.; Frost, J. Benzene-Free Synthesis of Adipic
Acid. Biotechnol. Prog. 2002, 18, 201−211.
(32) Jimenez, N.; Curiel, J. A.; Reveron, I.; de Las Rivas, B.; Munoz,
R. Uncovering the Lactobacillus plantarum WCFS1 gallate decarbox-
ylase involved in tannin degradation. Appl. Environ. Microbiol. 2013,
7
(
9 (14), 4253−63.
33) Ebenau-Jehle, C.; Mergelsberg, M.; Fischer, S.; Bruls, T.;
̈
Jehmlich, N.; von Bergen, M.; Boll, M. An unusual strategy for the
anoxic biodegradation of phthalate. ISME J. 2017, 11, 224−236.
(34) Junghare, M.; Spiteller, D.; Schink, B. Anaerobic degradation of
xenobiotic isophthalate by the fermenting bacterium Syntropho-
rhabdus aromaticivorans. ISME J. 2019, 13, 1252−1268.
(35) Liu, Y.; Gong, R.; Liu, X.; Zhang, P.; Zhang, Q.; Cai, Y. S.;
Deng, Z.; Winkler, M.; Wu, J.; Chen, W. Discovery and character-
ization of the tubercidin biosynthetic pathway from Streptomyces
tubercidicus NBRC 13090. Microb. Cell Fact. 2018, 17 (1), 131.
(36) Hayakawa, H.; Motoyama, K.; Sobue, F.; Ito, T.; Kawaide, H.;
Yoshimura, T.; Hemmi, H. Modified mevalonate pathway of the
archaeon Aeropyrum pernix proceeds via trans-anhydromevalonate 5-
phosphate. Proc. Natl. Acad. Sci. U. S. A. 2018, 115, 10034−10039.
(37) Wieser, M.; Fujii, N.; Yoshida, T.; Nagasawa, T. Carbon
dioxide fixation by reversible pyrrole-2-carboxylate decarboxylase
from Bacillus megaterium PYR2910. Eur. J. Biochem. 1998, 257, 495−
9
(
.
38) Bailey, S. S.; Payne, K. A. P.; Fisher, K.; Marshall, S. A.; Cliff,
M. J.; Spiess, R.; Parker, D. A.; Rigby, S. E. J.; Leys, D. The role of
conserved residues in Fdc decarboxylase in prenylated flavin
mononucleotide oxidative maturation, cofactor isomerization, and
catalysis. J. Biol. Chem. 2018, 293, 2272−2287.
(
39) Ferguson, K. L.; Arunrattanamook, N.; Marsh, E. N.
Mechanism of the Novel Prenylated Flavin-Containing Enzyme
2
878
https://dx.doi.org/10.1021/acscatal.0c05042
ACS Catal. 2021, 11, 2865−2878